2016-10-25
The Problem: Physical Limits & Need for Automation = Complex Interfaces
Contents
I'll show you..
- What is the cloud? What for?
- Solution case studies.
The Cloud: A Metaphor
Serviced Apartments..
- You can't afford the entire building.
- You don't NEED the whole building.
- But you want an apartment, and you want some control over what you do there and who visits.
- You also don't want people in the other apartments ruining your experience.
- You just need it for a few weeks. And more rooms this weekend for a party!
The Cloud
Some clouds..
Some common and confronting aspects of clouds..
- Key pairs not passwords
- Security groups
- Storage 'volumes'
- 'Object' storage
Commercial Clouds
| Pros | Cons |
|---|---|
| 'Infinite' Power | Often Expensive |
| On-Demand | Easy to be irresponsible |
| Upgrades | Can get 'locked in' |
| Reliable | Data soverignity worries |
| Competition | … |
| No Procurement, Ops, Space | … |
Data soverignity: Microsoft Azure
"European Union (EU) data protection law regulates the transfer of EU customer personal data to countries outside the European Economic Area (EEA).."
…
"Microsoft has invested in the operational processes necessary to meet the exacting requirements of the Model Clauses for the transfer of personal data to processors."
(https://www.microsoft.com/en-us/TrustCenter/Compliance/EU-Model-Clauses)
Non-Commercial Clouds
| Pros | Cons |
|---|---|
| Cost Effective (merit)! | Booked out |
| Great for long-running services | Unreliable? |
| On-site | … |
Over-all recommendation: They're complimentary solutions. Long-running compute, eg web hosting and services is quite good on a non-commercial cloud. 'Bursty' on AWS.
Case Studies
Genomics Virtual Lab: Bioinformatics Gateway Drug
Stuart's Nature Paper + App: Perfect Cloud Exemplar
Monash Bioinformatics Platform's Day to Day Compute and Storage
